The SMMUv3 architecture permits caching of data structures deemed to be "reachable" by the SMU, which includes STEs marked as invalid. When transitioning an STE to a bypass/fault configuration at init or detach time, we mistakenly elide the CMDQ_OP_CFGI_STE operation in some cases, therefore potentially leaving the old STE state cached in the SMMU.
This patch fixes the problem by ensuring that we perform the CMDQ_OP_CFGI_STE operation irrespective of the validity of the previous STE. Cc: Robin Murphy <robin.mur...@arm.com> Reported-by: Eric Auger <eric.au...@redhat.com> Signed-off-by: Will Deacon <will.dea...@arm.com> --- drivers/iommu/arm-smmu-v3.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/drivers/iommu/arm-smmu-v3.c b/drivers/iommu/arm-smmu-v3.c index 47f52b1ab838..91fdabdb4de6 100644 --- a/drivers/iommu/arm-smmu-v3.c +++ b/drivers/iommu/arm-smmu-v3.c @@ -1085,8 +1085,11 @@ static void arm_smmu_write_strtab_ent(struct arm_smmu_device *smmu, u32 sid, dst[1] = cpu_to_le64(STRTAB_STE_1_SHCFG_INCOMING << STRTAB_STE_1_SHCFG_SHIFT); dst[2] = 0; /* Nuke the VMID */ - if (ste_live) - arm_smmu_sync_ste_for_sid(smmu, sid); + /* + * The SMMU can perform negative caching, so we must sync + * the STE regardless of whether the old value was live. + */ + arm_smmu_sync_ste_for_sid(smmu, sid); return; } -- 2.1.4 _______________________________________________ iommu mailing list iommu@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/iommu